Monitor websites, APIs, and web pages around the clock. The moment something shifts - you're the first to know.
No credit card required
Change detected
Slack and email notifications are dispatched with screenshot and DOM diff context.
Plain English; Claude decides.
Keyboard-first j/k/u/n/e.
Structured URL & rule diffs.
Detect CDN-by-region drift.
One number replaces threshold tuning.
Features
Everything you need to stay ahead of changes - designed for developers and teams.
Write a rule in plain English ("alert me when the Pro plan price drops below $50"); Claude reads the diff and decides whether each change matches your intent.
A keyboard-first changes inbox. Unreviewed / Useful / Noise / Archived columns with j/k navigation, u/n verdicts, e to archive. Noise verdicts mine repeated patterns and propose mask regexes.
First-class monitor type. We diff added and removed URLs across the sitemap (recursing into sitemap-index documents), not the raw XML, so re-serialisation never trips an alert.
Per-user-agent rule diffs, sitemap directive changes, host directive changes. Catch unintentional Disallow rules before they tank organic traffic.
Monitor a URL from multiple regions. When the regional hashes diverge consistently and each region is individually stable, we classify the change as geo-targeted and tell you which regions disagree.
Every change gets a 0-1 score that combines visible-text diff, structural diff, change class, SEO drift, JSON-Patch density, section density, keyword intent, and reword penalty. Set 'fire when confidence > 0.5' once instead of guessing thresholds per monitor.
Extract prices, stock counts, or any numeric tracked key and render a 7/30/90-day time series with first/last/min/max/delta on the monitor detail.
Field-level JSON-Patch for API responses, section-anchored diffs for long HTML, SEO drift for head tags. The line diff is still there; the structured artefacts are what the UI and alert rules reason over.
Screenshot-based comparison detects pixel-level changes across your monitored pages.
Target specific elements. Ignore headers, footers, and ads - watch only what matters.
Track JSON and REST endpoints with custom headers, authentication, and request bodies.
Route notifications to Slack, Discord, email, or webhooks - per monitor, your rules.
Poll as often as every 10 seconds. Time-sensitive pages get the attention they deserve.
Full REST API for programmatic monitor management. Build dashboards and automations.
Use cases
Pick the persona closest to your job and see the features that solve it.
Sitemap drift, robots.txt rule changes, title and meta tag diffs, JSON-LD schema watches.
WCAG regressions tracked against an attested baseline, AI a11y impact classification, client-ready VPAT evidence.
Tracked prices and stock as time series with charts and alert rules that fire on threshold crossings.
API contract diffs with field-level JSON-Patch, GitHub commit attribution, confidence-scored alerts, multi-region geo classification.
Section-anchored diffs on long documents (ToS / privacy / docs) and signed change-verdict links direct from alert emails.
How it works
Paste any website, API endpoint, or page URL. Use CSS selectors to target specific elements.
Our infrastructure checks your URLs on schedule - text diffs, visual screenshots, or API tracking.
The moment a change is detected, you receive an alert with a detailed diff.
Pricing
Start free and scale with your needs.
Monitoring playbooks
OnChange combines website change detection, visual diff monitoring, SEO monitoring, API tracking, and accessibility scans in one workflow. Use it to protect the pages and endpoints that customers, search engines, and internal teams rely on.
FAQ
Quick answers for teams comparing OnChange with uptime tools, visual diff tools, and manual QA checks.
A website change can be edited copy, a redesigned section, a changed price, a removed CTA, a different API response, a new accessibility violation, a metadata update such as a title tag or canonical URL, a sitemap.xml URL addition or removal, or a robots.txt rule change.
Use faster checks for revenue-critical pages, pricing, APIs, and launch pages. Daily or weekly checks are usually enough for evergreen content, legal pages, and documentation.
Yes. Target a specific element with CSS selectors, ignore headers or ads, mute volatile JSON paths or sections with regex masks, set a confidence threshold on the composed signal score, or write a plain-English rule like "alert me when the Pro plan price drops below $50" and let Claude decide whether each diff matches.
Yes. Uptime monitoring tells you whether a page responds. Change monitoring tells you whether the page still says, shows, and returns the right thing.
The /changes view is a four-tab inbox - Unreviewed, Useful, Noise, and Archived - with keyboard shortcuts (j and k navigate, u marks useful, n marks noise, e archives). Marking changes as noise feeds an aggregated suggestion engine that mines repeated patterns and proposes regex masks to silence the same noise on future checks.
When you configure multiple regions on a monitor, every check runs from each region via a regional proxy. The classifier compares the regional hashes against each other and against the primary; when regions disagree consistently it tags the change as geo-targeted so you can spot CDN A/B rollouts, feature flags by source IP, and price experiments by country.